talk-data.com

We’ve built North America's most capable pharmacy agents using LangChain and Elasticsearch, securing over $1M in pre-seed funding and serving pharmacy chains with over a thousand stores. I was at the Google Montreal office 2 weeks ago sharing my experiences at a genAI panel. I’d love to give a lightning talk at the Elastic and LangChain NYC Meetup on building high-accuracy agents for high-risk industries. Healthcare is a tough industry but LangChain helped us tremendously. Our journey through due diligence and navigating legacy sectors could offer valuable insights.

Introduce LangGraph; Demo notebook of a RAG agent built w/ LangGraph; Demo some upcoming products from LangChain

In this talk, I will demonstrate how we leverage LangChain embedded in OBTO's application development platform to create developer agents capable of building and updating end-to-end web applications. OBTO introduces an innovative approach to structuring applications, which, combined with effective prompt engineering and fine-tuned models, significantly enhances the accuracy of these agents. Our platform ensures seamless communication across the tech stack using a unified language, and it incorporates flexible ways to optimize and manage the performance of these interactions. Additionally, OBTO enables continuous feedback loops to refine and improve the agent setups, ensuring a robust and efficient development process.

Discover how evaluations and LangSmith helped improve its performance and effectiveness.

Earlier this month, my band did a run of shows in the Shanghai area, and there were some press clippings in Chinese that I needed to translate. I built an example application that translates the text of these clippings from Chinese to English using Doctran and uses vector search in Elasticsearch to learn more about the translated documents.